Geological Identification Based On K-Means Clustering Of Data Tree Of Shield Tunneling Parameters

With the development of the country,the country continuously increases the investment in infrastructure construction.The construction of expressways,railways,subways and other basic transportation has broken down the barriers of national economic development and urban construction,injected vitality into national economic development,and provided convenience for the life of urban residents.Shield tunneling machine has become the main machinery for traffic construction in China because of its high efficiency,low disturbance and adaptability to various geological conditions.Before the shield method is used for construction,geological conditions need to be explored in the early stage,geological information should be mastered in order to select the appropriate shield machine type,and reasonable schemes should be provided to control the shield machine construction.However,due to the complexity and variability of geological information and the difficulty of accurate prediction,construction will face great risks and even lead to serious safety accidents when encountering unfavorable geology.Real-time and accurate geological identification and prediction is an important guarantee for the safe construction of shield machine,which is also an important problem in shield technology.Therefore,this dissertation uses data mining method to identify and analyze geological categories based on shield tunneling parameters:(1)The main parameters of shield machine are theoretically analyzed,and the influence of geological characteristics on shield tunneling parameters is analyzed.The correlation between shield tunneling parameters and geological categories is analyzed so as to select suitable shield tunneling parameters for geological identification.(2)Based on real-time shield tunneling parameters,a geological identification method based on unsupervised data tree K-Means clustering is proposed.The data tree is used to optimize the K value in the K-Means clustering algorithm,and then the geological identification clustering algorithm based on the main tunneling parameters is constructed.Then based on the field construction data,a simulation experiment is carried out,and the correct rate of geological identification reaches 100%.The results show that the method can accurately identify geological types and provide certain decision support for shield machine to implement effective parameter control.(3)The BP neural network is trained with the training sample data under single geological condition and the training sample data under composite geological condition,and two models are obtained.The two models are used to predict the earth pressure of the same test set respectively,and the errors of the prediction results are compared.The results show that the prediction model obtained from the training sample data under a single geological condition has higher accuracy.It is further proved that geological identification can provide effective support for setting or adjusting tunneling parameters.